Yi Zhangandispobock 2998c4bdf4 [optimize] fuse renormalize into moe_topk_softmax (#7744) ...

Co-authored-by: ispobock <ispobaoke@gmail.com>

2025-07-03 12:42:44 -07:00

..

bench_awq_dequant.py

…

bench_cutlass_mla.py

…

bench_dsv3_fused_a_gemm.py

…

bench_dsv3_router_gemm.py

…

bench_fp8_blockwise_gemm.py

refactor apply_w8a8_block_fp8_linear in fp (#6545)

2025-05-29 00:15:11 -07:00

bench_fp8_blockwise_group_gemm.py

…

bench_fp8_gemm.py

…

bench_int8_gemm.py

…

bench_lightning_attention_decode.py

[Fix] use torch.cat instead of torch.concat to prevent entering the Autograd backends. (#4466)

2025-03-16 00:02:47 -07:00

bench_moe_align_block_size.py

…

bench_moe_ep_post_reorder.py

…

bench_moe_ep_pre_reorder.py

…

bench_moe_fused_gate.py

…

bench_moe_silu_and_mul.py

…

bench_moe_topk_softmax.py

…

bench_per_tensor_quant_fp8.py

…

bench_per_token_group_quant_8bit.py

…

bench_per_token_quant_fp8.py

…

bench_qserve_w4a8_gemm.py

…